Challenge
 

A leading online assessment and proctoring service provider required a sophisticated real-time reporting system to enhance their data streams. The goal was to improve transparency and efficiency in proctoring processes, enabling clients to make informed decisions based on real-time data insights.

Following were the key requirements:

• Real-Time Reporting Needs: The client needed to capture and report on various metrics including authorization speeds, exam quality, proctor performance, and candidate issues in real-time.

• Data Integration: The system needed to handle diverse data sources including MongoDB, Postgres, MSSQL, MariaDB, Oracle, MySQL, and SNS events, requiring seamless integration and processing.

• Performance Monitoring: Ensuring service health monitoring, including network speed, CPU, memory utilization, and API request success rates, was crucial for self-evaluation and system reliability.

Approach

Following were the key components of the system:

• Data Streaming and Storage: Implemented Apache Kafka for robust data streaming and a secondary ingestion tool to handle the high volume of data. Delta Lake was used for efficient data storage close to the source.

• Reporting Tools: Utilized Postgres for sub-second level reporting. Jasper Reports was employed for user-facing reports, while Metabase was used for internal reporting.

• Comprehensive Data Sources: Integrated various data sources including MongoDB, Postgres, MSSQL, MariaDB, Oracle, MySQL, and SNS events to ensure a holistic view of the system's performance.

Xponentium Impact
 

• Exam Quality and Proctor Performance: Enabled detailed tracking of violations, response coverage, and proctor interaction with candidate feeds, leading to improved exam integrity and proctor effectiveness.

• Candidate Issues Tracking: Improved tracking of active candidates, proctors, and authorizers, monitoring connection drops, and internet speeds, which helped in addressing technical issues promptly.

• Service Health Monitoring: Provided insights into network speed, CPU, memory, and disk utilization, as well as average response times and API request success rates, enhancing overall system reliability and performance.

• Authorization Performance: Enhanced reporting on authorization speeds and quality, significantly reducing false positives and false negatives while improving queue monitoring.